Historical floods in the Dutch Delta R. Glaser, H. Stangl

To cite this version:

R. Glaser, H. Stangl. Historical floods in the Dutch Rhine Delta. Natural Hazards and Earth System Sciences, Copernicus Publ. / European Geosciences Union, 2003, 3 (6), pp.605-613. ￿hal-00299077￿

HAL Id: hal-00299077 https://hal.archives-ouvertes.fr/hal-00299077 Submitted on 1 Jan 2003

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Natural Hazards and Earth System Sciences (2003) 3: 605–613 © European Geosciences Union 2003 Natural Hazards and Earth System Sciences

Historical floods in the Dutch Rhine Delta

R. Glaser and H. Stangl University of Heidelberg, Department of Geography, INF 348, 69120 Heidelberg, Received: 23 September 2002 – Revised: 6 February 2003 – Accepted: 12 February 2003

Abstract. Historical records provide direct information On a global as well as on a regional scale, this relatively about the climatic impact on society. Especially great natural short phase of up to about 200 years is still the only com- disasters such as river floods have been for long attracting the monly recognized scale for characterizing climate and flood attention of humankind. Time series for flood development patterns. The question arises on whether the changes within on the Rhine branches , / and IJssel in the this period are part of a long-term natural development or a Dutch Rhine Delta are presented in this paper. In the case of sign of the human impact on climate. Furthermore we have the Waal it is even possible to compare historical flood fre- to ask which dimension these changes have in a historical quencies based on documentary data with the recent devel- context. opment reconstructed from standardized instrumental mea- Up to a certain point past climate research can provide an- surements. In brief, we will also discuss various parameters swers to these questions. Historical records, in various re- concerning the structure of the flood series and the “human gions of Central Europe dating back more than 1000 years, dimension” of natural disaster, i.e. the vulnerability of soci- provide direct information about the climatic impact on soci- ety when facing natural disasters. ety. Their high temporal resolution is of special value for the examination of natural disasters such as river floods. Besides extensive reconstructions of singular historic flood events (Corsten, 2001; Munzar, 2001; Tetzlaff et al., 2001), vari- 1 Introduction ous publications present long time series of flood frequen- Flood events such as the one along the Danube and Elbe cies (Brazdil´ et al., 1999; Portge¨ and Deutsch, 2000; Sturm River in the summer of 2002 regularly cause discussions et al., 2001). Figure 1 illustrates the findings of several Cen- about the influence of global warming. There is evidence tral European rivers. In this paper we will discuss the flood that the frequency of strong precipitation events generally in- development along the Rhine branches Waal, Nederrijn/Lek crease. Until 2050, most climate change scenarios predict an and IJssel in the Dutch Rhine Delta (Fig. 2). Time series dat- increase in the average annual discharge of about 10% north ing back to AD 1350 will be presented as well as a concise of the Alps (IPCC and McCarthy, 2001). In order for us to discussion on their underlying control mechanisms. Further- be able to create and validate appropriate models we need to more, the human dimension of flood events will be briefly know and understand climatic development of the past such reviewed. as the frequency of floods. Since the late 18th century, flood patterns in Central Eu- rope have been evaluated on the basis of official and stan- 2 Flood data and its interpretation dardized measuring data. In various regions climate change seems to be a factor for the intensification of natural disas- 2.1 Flood observation in the Dutch Rhine Delta ters both in number and extent. For the Rhine catchment area, winter precipitation has increased approximately 30%. Since the beginning of our historic time, climate and its ex- These changes go along with a corresponding increase of an- tremes have been attracting the attention of humankind. Peo- nual flood discharge from 1891 on – for example at the gaug- ple wrote documents and recorded early instrumental read- ing station Cologne (Mendel et al., 1997). ings. In Western and Central Europe, climate observations in historical documents have existed since the Carolingian Correspondence to: R. Glaser era. After Pfister (2001), looking at the amount and conti- ([email protected]) nuity of this material, we can divide the past 1250 years into 606 R. Glaser and H. Stangl: Historical floods in the Dutch Rhine Delta

Lower Rhine

Pegnitz

Main

Danube 31-year running mean, standardized

1300 1400 1500 1600 1700 1800 1900 2000 Year

Fig. 1. 31-year running mean of historical floods on various European rivers (Glaser and Stangl, revised).

Table 1. Periods of climate observation in Central Europe (after Pfister, 2001)

from temporal resolution availability source type

∼ AD 700 - (yearly) sporadic document ∼ AD 1250 seasonal almost uninterrupted document ∼ AD 1500 monthly (daily) almost uninterrupted document ∼ AD 1680 daily sporadic instruments, individuals ∼ AD 1860 daily (intra-daily) uninterrupted instruments, officials

five periods with gradually increasing source availability and parameters, national and international measuring networks temporal resolution (Table 1). were established, providing objective and standardized data In the area of the Dutch Rhine Delta we may assume suf- sets. At the gauging station along the Waal daily ficient source availability on from about AD 1350. Histori- acquisition of water level data was started as early as 1770, cal records such as the “Cameraarsrekeningen van Deventer” following some large flood events in the preceding years. and the “Stadsrekeningen van ” provide valuable in- Today online data banks make this information available formation over many centuries. The financial consequences even to the general public. River measuring data for the of river floods concerning dike protection and repairs are es- have been compiled by the “Rijkswaterstaat” pecially mentioned in these accounts. Monasteric chronicles (http://www.waterbase.nl). or weather diaries written by individuals give further infor- mation about the extents and often even about the causes of 2.2 Data evaluation and interpretation flood events. Thanks to compilations edited by authors like Weikinn While dealing with historical data, one has to be aware of the (1958–1963), Gottschalk (1971–1977), and Alexandre subjective nature of written sources. The reason for this is the (1987), lots of historical sources are easily available today. fundamental influence of man’s individual perception. Gen- To provide modern data handling, a graphical presentation, erally speaking, separating subjective from objective compo- and the possibility for statistical analysis, data banks like nents of historical sources requires historical methods of data the “HISKLID” (http://www24.brinkster.com/hisklid/index. evaluation and interpretation, including critical source anal- htm) have been developed. ysis and hermeneutics. Where available, interregional com- In the course of enlightenment and the invention of new parison by means of source synopsis is of special benefit for technical devices we notice the emergence of the first quan- the acquisition of reliable data. Figure 3 shows a scheme for titative data sets. They are provided by individual instru- the critical evaluation and interpretation of historical docu- mental measurements and flood marks on bridges, houses, or ments. city ports. Finally, for the regular acquisition of hydrologic The historical methods help us to judge the reliability of R. Glaser and H. Stangl: Historical floods in the Dutch Rhine Delta 607 The Dutch Rhine Delta (present situation)

Type of Source: chronicles, annals, town chronicles, diaries, calendars

Specification: primary/ secondary numerical/ descriptive

North Sea Verification: bibliographical level environmental level lifedataoftheauthor(s) general Zeitgeist education perception of climate professional career and weather general circumstances of life rational identification patterns level of mental perception THE NETHERLANDS intention and motivation knowledge horizon Ijsselmeer eyewitness/ contemporaries objectifiable reality

I Amsterdam js Deventer historical and scientific facts, source synopsis s e l technical level Den Haag calendar adaptation, reference to a certain locality/ regionalisation Nede Lek r rij n Arnhein Nijmegen source evaluation - climate information Waal

L Quantification: linguistic level statistical level ow er R hi ne deduction of semantic profiles statistical basic parameters classification frequency counts Essen

Antwerpen indices time series analysis Dusseldorf calibration Gent quantitative climate time series GERMANY Cologne Brussels BELGIUM . Fig. 3. Scheme for source criticism (Glaser, 2001).

Fig. 2. The Dutch Rhine Delta. limiting value for classifying gauge levels as flood events. historical data. One next step is their quantification. There- 2.3 Analyzing flood development fore understanding hydrological processes is as fundamental For the graphical presentation of the flood series we chose as statistical tests and multiple statistical methods. A com- 31-year running frequencies on the basis of the hydrological mon procedure is the derivation of indices. Levels of magni- year (starting on 1 November). Though this measure exhibits tude (e.g. duration of the flood or buildings damaged by the a relatively poor filtration effect (Schonwiese,¨ 1992), it meets water), expressed in the records, are correlated with numeri- the needs of various geoscientific approaches with respect cal methods. Taking into account the descriptions about the to the commonly used 30-year time span for climatological flood consequences on other natural phenomena, for instance phases. the phenological status or soil conditions, Sturm et al. (2001) Due to the multitude of uncertainties concerning the devel- such developed a scheme for the intensity classification of opment of the processes controlling flood evolution, and due historical floods (Table 2). to the lack of a comprehensive model describing the complex A crucial problem is the combination of historical data interactions between these processes, the discussion of the based on the description of flood damages, and discharge time series mainly refers to some selected points in a quali- data based on water level measurements into a single com- tative way. parable series. Sometimes we even have to deal with a gap For an elementary quantitative discussion, t-test analysis of several decades between historical and instrumental data. seems to be appropriate (Schonwiese,¨ 1992). For a specific For this case, Sturm et al. (2001) have developed a method year we considered the flood frequencies of the 30 preceding based on the deviation of the monthly maximum discharge and the 30 subsequent years to be random samples. Gaussian from the mean discharge of a reference period with respect distribution for each random sample was verified by using the to the standard deviation. Kolmogorov-Smirnov test (Schonwiese,¨ 1992). As a t-test In this paper, we will discuss a long time series for the estimator we get Waal. Historical records were available up to AD 1800, √ |¯a − b¯| n whereas daily instrumental measurements were initiated in ˆ = t q (1) 1770. For the intersection period of 31 years we compared 2 + 2 sa sb written information and river level data. Based on the neces- sary condition that both kinds of data should give information with n = 30 (number of elements in the random samples) 2 2 about the same number of river floods, we thus determined a and sa , sb variances of the random samples a and b. 608 R. Glaser and H. Stangl: Historical floods in the Dutch Rhine Delta

Table 2. Intensity classification of historical floods (Sturm et al., 2001)

Level Classification Primary Indicators Secondary Indicators

1 smaller, regional flood little damage, e.g. fields and gardens close short flooding to the river, wood supplies that were stored close to the river are moved to another place 2 above average, or supra-regional, damage to buildings and constructions flood of average duration; severe damage flood related to the water like dams, weirs, to fields and gardens close to the river, footbridges, bridges and buildings close to loss of animals and sometimes people the river, like mills, etc.; water in buildings 3 above average, or supra-regional, severe damage to buildings and constructions duration of flood: several days or weeks; flood on a disastrous scale related to the water, i.e. dams, weirs, severe damage to fields and gardens close footbridges, bridges and buildings close to to the river, extensive loss of animals and the river, like mills etc.; water in buildings. people; morphodynamic processes like In part, buildings are completely destroyed sand sedimentation cause lasting damages or torn away by the flood and change the surface structure

Table 3. Basic types of flood events in Central Europe (after Brazdil´ et al., 2002)

flood type spatial scale meteorological causes/ synoptic situation

flash flood (due to short but local (but often with disastrous connected with heavy precipitation during intensive downpours/ cloudbursts) impacts) convectional storms

flood due to long-lasting continuous rainfalls related to supra-regional flood due to snow-melt larger territorial units synoptic patterns (such flood due to ice damming as cyclones or troughs)

Waal 30 20 10 0 Nederrijn/ Lek 30 20 10 0 IJssel 30

31-year running frequency 20 10 0 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 Year

Fig. 4. Historical floods in the Rhine Delta. R. Glaser and H. Stangl: Historical floods in the Dutch Rhine Delta 609

− = For 2n 2 58 DOFs (Degrees of Freedom) and an al- Table 4. Flood peak values on the Waal, hydrological years pha level of 1% we get a significance level of t(58; 0.01) = 1770–1800 (source for Nijmegen gauge data: Rijkswaterstaat, http: 2.392. Larger estimators, especially very high and narrow //www.waterbase.nl) peaks in the graph of the estimator function give indication for fracture points in the flood time series, which may be starting-points for further investigations. rank date Nijmegen gauge registered in (cm above NAP)* historical documents? 2.4 Flood typology in Central Europe 1 21.02.1799 1378 X Brazdil´ et al. (2002) characterise four basic types of flood 2 27.01.1781 1367 X events according to their meteorological causes (Table 3). 3 08.02.1799 1367 X Flash floods are of minor importance for the Dutch Rhine 4 03.01.1784 1344 X Delta, whereas the three flood types concerning larger terri- 5 Dez. 1769 > 1265 supposed X torial units have occurred in changing frequency in the past 6 14.02.1795 1265 – either in pure form or overlapping one another. 7 25.01.1789 1250 X 8 29.01.1778 1239 Flooding due to ice damming may be considered as a spe- ··· cial case. During a long period of very low temperatures, es- ··· pecially river reaches with small flow velocities can be cov- ··· ered by ice. This condition is fulfilled in the generally slowly dropping Rhine branches, especially in the transition regions to the tide. From there on, the ice cover could subsequently be extended upstream. A sudden warming, in Central Eu- ropean winters often combined with precipitation from At- lantic air masses, could then lead to the piling up of ice clods few floods did not always occur simultaneously. The reason at baffles in the river, such as sand banks. for this will be discussed in the following chapter. As explained above, t-test analysis may help us to find the 3 Time series and their structure fracture points in the flood series (Fig. 5). The periods within which t-test estimators of all the three Rhine branches sur- 3.1 Historical floods along Waal, Nederrijn/Lek and IJssel pass the significance level may be of special interest with respect to large scala phenomena. For example, this is the Close to the German-Dutch border near Lobith the bifurca- case between AD 1385 and AD 1413. While flood risk sig- tion of the Rhine into the Waal and the Pannerden Canal nificantly increases along the Waal and Nederrijn/Lek within marks the commencement of the Rhine Delta. About ten this phase, we have significantly less flood reports for the kilometers further downstream the latter itself branches out IJssel in the years following about AD 1410. into the IJssel and the Nederrijn, in its lower part, after cross- ing the Amsterdam-Rhine-Canal near bearing the name Lek. The evolution of the Rhine Delta dur- 3.2 Adhering instrumental data to the Waal series ing the Holocene has been described by Brunnacker (1978), Woude (1981) and Czaya (1981). Figure 4 shows the 31-year running frequencies of histor- The comparison of documentary and instrumental data for ical floods on the Waal, Nederrijn/Lek and IJssel (levels 1, the Waal exhibits a good conformity (Table 4). The five 2, and 3, according to the scheme presented in Table 2). largest flood peak values have been referred to in historical On from about 1350 AD, in the case of the Waal proba- documents as well. The flood event ranking at number six bly a few decades later, we may assume a sufficient source has not been described, while number seven is included in availability. First of all, the running frequencies exhibit sig- the records once more. The marginal difference of 15 cm be- nificant changes in flood frequencies. For example, along tween the latter 2 events might already be due to intra-day Nederrijn/Lek and Waal, a notable peak is situated around fluctuations. Thus we may state that within the time span AD 1420, whereas relatively few floods occurred in the sec- 1770–1800 in fact large river discharges have been described ond half of the fifteenth century. Apparently, medium-term as floods whereas normal ones have not. A limiting value of increases and decreases within a range of 30 to 100 years 1257,5 cm (mean of the gauge levels of the flood events on were quite common. This finding is remarkable as these de- rank six and seven) thus seems to be appropriate for classi- viations occurred within the period of “natural climate”, i.e. fying subsequent events as floods. On this base we classified in a time, when humankind did not yet cause global warming. the daily water levels at the Nijmegen gauge for the years The hydrological regime of all the three Rhine branches is 1801–2001. The relocation of the gauge station in the year almost exclusively controlled by the discharge of the Lower 1980 has been considered by an appropriate compensation. Rhine. Yet, periods with frequent floods and periods with Figure 6 shows the resulting flood series. 610 R. Glaser and H. Stangl: Historical floods in the Dutch Rhine Delta

Waal t (58; 0,01) 4

Nederrijn/ Lek 4

IJssel 4

0 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 Year

Fig. 5. t-test estimators (30 years preceding/30 years following) of flood data.

25

documentary data 20 Nijmegen gauge

15

10

5 31-year running frequency

0 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900 1950 2000 Year

Fig. 6. Flood development on the Waal since AD 1350 (source for Nijmegen gauge data: Rijkswaterstaat, http://www.waterbase.nl).

4 Discussion It is important that these control mechanisms must not be understood unilaterally. Floods themselves influence the 4.1 Parameters concerning flood series structure of the river basin as well as the use of preventive measures and human response, thus exhibiting the main char- Flood patterns are subject to a multitude of influences – not acteristics of a feedback operation. only to climatic parameters but also to other geophysical as- pects such as the morphometric structure of the river basin or Furthermore, Fig. 7 illustrates a fundamental problem of to social aspects (Fig. 7). The problem of the source effect arranging historic and recent flood data as well as of inter- has already been discussed in Sect. 2.2. Critical source anal- preting flood series in general. While documentary records ysis is applied in order to reduce this effect to a minimum. are in fact subject to all the parameters illustrated, instrumen- The other aspects illustrated may be of various importance tal data are neither concerned by human flood exposure nor for different river basins. E.g., while land-use changes (such influenced by source effects. as woodland clearance) may have dramatic effects in small When giving a limiting value for an intersection period river basins, their effect on the Rhine system with its large of documentary and instrumental data, we therefore have to catchment area seems to be of minor importance (Kwadijk, be aware, that this value has not necessarily been constant 1993). in the time preceding the instrumental period, for which we R. Glaser and H. Stangl: Historical floods in the Dutch Rhine Delta 611

Availability Perception

SOURCES

Housing River Bed

EXPOSURE RIVER BASIN FLOOD SERIES

0,2 0 Hazard Catchment Area Management

ATMOSPHERE

0 0

Climate Weather

Fig. 7. Parameters concerning historic

0 0 0 0 flood series.

classified the floods due to their effects (Table 2). Knowledge ment around the year of “St. Elisabeth’s flood”, to an earlier about possible changes would be of fundamental importance storm flood event already altering the river flow patterns – for quantifying the other effects illustrated, such as climatic Gottschalk (1975) describes two more disastrous events for changes. the years 1404 and 1409 – or for other reasons may not be In addition, developing appropriate filter techniques seems stated for sure at this point. to be of fundamental importance for the comprehensive inter- After centuries of further storm floods and silting up of pretation of flood development. Nederrijn/Lek and IJssel, water distribution between the In the following, we will present some selected aspects to Rhine branches was finally stabilized in the course of the illustrate the complex interactions characterizing flood devel- 18th century, especially by the construction of the Panner- opment in the Dutch Rhine Delta. den Canal (Urk and Smit, 1989; Ven et al., 1995). Large scale river regulations and excavations of sand 4.2 Floods and the changing structure of the river basin banks were among the elaborative measures implemented in the 19th and 20th centuries, both aiming to improve protec- Changes in the catchment area as well as changes along the tion from floods and to ease navigation (Tummers,¨ 1994). river itself may be of great influence on the frequency and extents of river floods. In the Netherlands, dike construction 4.3 Floods and atmospheric patterns – followed by sand sedimentation and damming up of river water – was an important factor (Tol and Langen, 2000; Ven Of course, there is a significant correlation between floods et al., 1995). This can be easily understood in the case of and the variability of our climate. On different time scales, icefloods (which formerly occurred more frequently, see be- regular and irregular climatic variations define the conditions low). On the other hand, nowadays, dikes are indispensable under which specific weather situations (pressure, temper- means for the protection of the human settlements close to ature, precipitation) lead to the formation of flood events. the rivers. Glaser (2001) presents a reconstruction of the temperature’s Furthermore, historic storm floods caused great losses of and precipitation’s development within the past millennium land in the area of river mouths. The displacement of the ero- based on documentary records. sion base was followed by significant redistribution of Rhine Our examination of historic flood data led to the suppo- water between the various branches. Special attention has sition that during the second half of the 17th century, that been given up to this day to the catastrophic “St. Elisabeth’s is to say, during a climax period of the Little Ice Age, up flood” in AD 1421 (Ven, 1996; Gottschalk, 1975). to about 70% of flood events in the area of the Rhine Delta Yet, as mentioned above, a fracture point for the flood might have been in connection with icing (often in combina- series is supposed to be situated at least one decade ear- tion with precipitation and/ or snow melt). That ice floods lier. Whether this is due to an opposing climate develop- are of negligible importance nowadays is not only due to cli- 612 R. Glaser and H. Stangl: Historical floods in the Dutch Rhine Delta matic effects, but also to human interventions such as dense high temporal and spacial resolution proved to be a special navigation and warming up of the river by industrial cooling benefit. water (CHR, 1978; Ven et al., 1995). It became clear that river floods in Central Europe have always happened, and that the appearance of these disasters 4.4 Floods and society drastically changed during the past 700 years. By looking at these phases, we have to assume that during some periods The impact floods have on society is a main aspect of histor- and in some regions floods occurred more frequently than ical flood evidence. Important features are: during the past two centuries. On the other hand, in some – the number of people living in settlements close to the regions, recent flood development seems to be unique within river the reconstruction period. This might be in connection with the phenomena of man-made global warming. – the strategies developed for managing flood hazards, for But we have to be careful with such a supposition. The instance warning and protection systems as well as mea- reason for this is that, in addition to atmospheric parameters, surements taken after the flood flood evolution is subdued to a multitude of parameters. Of special interest is the “human dimensions” of floods. Histori- Glaser et al. (2002) drew a multi-layered picture of disaster- cal documents provide broad information about the historical perception, disaster-management and vulnerability towards perception and management of natural disasters. natural disasters for various European rivers. Severe short- In order to apply high-quality filtration techniques, both ages of food as well as problems with clear drinking wa- in the fields of modelling and of data acquisition, further re- ter supply during and until several weeks after a flood were search is necessary to get a comprehensive understanding of among the most serious problems. past flood development. We consider the presented time se- This “human dimension” of floods has been of great im- ries and their elementary discussion as a first step towards portance in the Netherlands. Tol and Langen (2000) give an this direction. extensive description of the course of human settlement and cultivation close to the river, which reached a considerable level as early as the Medieval Optimum. The authors sup- References pose connections not only with technical inventions such as the wind mill but also with an early discernment that coop- Alexandre, P.: Le climat en Europe en Moyen Age, Ed. de L’Ecole eration is for the benefit of society and with the early estab- des Hautes Etudes en Sciences Soc., Paris, 1987. lishment of central governments by the counts of Gelre and Brazdil,´ R., Glaser, R., Pfister, C., Antoine, J. M., Barriendos, M., Kleve. Camuffo, D., Deutsch, M., Enzi, S., Guidoboni, E., and Ro- Ven et al. (1995) draw a detailed picture of the history of drigo, F. S.: Flood events of selected rivers of Europe in the prevention, warning systems, counter measures, and disaster Sixteenth Century, in: Climatic Variability in Sixteenth Century relief operations. In many communities, storage of materials Europe and its Social Dimension, edited by Pfister, C., Brazdil,´ and works to be executed in the case of a disasters were di- R., and R. Glaser, Kluwer, , 239–285, 1999. Brazdil,´ R., Glaser, R., Pfister, C., and Stangl, H.: Floods in Europe rected by legislation. Out of various precursor organizations, – A Look into the Past, PAGES News, 10 (3), 21–23, 2002. in the time of the Batavian Republic, a country-wide, cen- Brunnacker, K.: Der Niederrhein im Holozan,¨ Fortschritte der Ge- trally organized system of river correspondence was devel- ologie Rheinland und Westfalen, 28, 399–440, 1978. oped. Furthermore, the authors describe donation activities CHR (Commission Internationale de l’Hydrologie du Bassin du organized by wealthy citizens in large towns in the 18th cen- Rhin) (Ed.): Le Bassin du Rhin, 3 vols., 1978. tury. Historical documents (Gottschalk, 1975) give evidence Corsten, S.: Anmerkungen zur mittelalterlichen Geschichte von Ue- of food deliveries by adjacent villages that already took place dorf, Bonner Geschichtsblatter,¨ 49/50, 2001. in medieval times. Finally, Tummers¨ (1994) refers to a con- Czaya, E.: Strome¨ der Erde, Aulis, Koln,¨ 1981. nection between the threats by the water and religious orders. Glaser, R.: Klimageschichte Mitteleuropas, 1000 Jahre Wetter, These aspects once more stress the fact that it was not only Klima, Katastrophen, Wiss. Buchgesellschaft, Darmstadt, 2001. the influence of flood development on society, but also the in- Glaser, R., Jacobeit, J., Deutsch, M., and Stangl, H.: Hochwasser¨ als historisches Phanomen,¨ in: Katastrophe oder Chance? Bay- fluence of society on flood development and perception that erische Akademie der Wissenschaften (Ed.), Hochwasser und played an important role in history. Thus, both historic and Okologie,¨ Pfeil, Munchen,¨ 2002. geoscientific aspects have to be taken into account. Glaser, R. and Stangl, H.: Climate And Floods In Central Europe Since A.D. 1000: Data, Methods, Results And Consequences, Surveys in Geophysics, revised. 5 Conclusions Gottschalk, M. K. E.: Storm floods and river floods in the Nether- lands, van Gorcum, Assen, 3 vols., 1971–1977. Historic documentary sources provide valuable information IPCC (Intergovernmental Panel on Climate Change) and Mc- about climate and its extremes, which we can transform into Carthy, J.: Climate Change 2001: Impacts, Adaptation and Vul- semi-quantitative information by applying historical and sta- nerability, contribution of Working Group II to the third assess- tistical methods. Critical source analysis is therefore of ma- ment report of the Intergovernmental Panel on Climate Change, jor importance. In the field of river floods, the comparably Cambridge Univ. Press, Cambridge, 2001. R. Glaser and H. Stangl: Historical floods in the Dutch Rhine Delta 613

Kwadijk, J.: The impact of climate change on the discharge of the in: ATV-DVWK Landesverband Bayern (Ed.): Hochwasser- River Rhine, Nederl. Geogr. Studies, 171, , 1993. Niedrigwasser-Risiken, Hirthammer, Munchen,¨ 5–22, 2001. Mendel, H. G., Fischer, P., and Hermann, A.: Hochwasser – Tol, R. S. J. and Langen, A.: A concise history of Dutch river floods, Gedanken uber¨ Ursachen und Vorsorge aus hydrologischer Sicht, Climatic change, 46 (3), 357–369, 2000. Bundesanst. fur¨ Gewasserkunde,¨ Koblenz, 1997. Tummers,¨ H. J.: Der Rhein, ein europaischer¨ Fluß und seine Munzar, J.: Summer floods in Central Europe in 1813 – an Analogy Geschichte, Beck, Munchen,¨ 1994. to Floods of 1997, Moravian Geographical Reports, 9 (2), 29–40, Urk, G. van and Smit, H.: The , Geomorphological 2001. Changes, in: Historical changes of large alluvial rivers, Western Pfister, C.: Klimawandel in der Geschichte Europas, Zur Entwick- Europe, edited by Petts, G. E., Moller,¨ H. and Roux, A. L., Wiley, lung und zum Potenzial der Historischen Klimatologie, OZG,¨ 12 Chichester, 1989. (2), 7–43, 2001. Ven, G. P. van de: Leefbaar laagland: geschiedenis van de water- Portge,¨ K. H. and Deutsch, M.: Hochwasser in Vergangenheit und beheersing en landaanwinning in Nederland, Matrijs, Utrecht, Gegenwart, in: Dynamik von Flora und Fauna – Artenvielfalt 1996. und ihre Erhaltung, Bayerische Akademie der Wissenschaften Ven, G. P. van de, Driessen, A. M. A. J., Wolters, W., and Wasser, J.: (Ed.): Pfeil, Munchen,¨ 139–151, 2000. Niets is bestendig..., De geschiedenis van de rivieroverstromin- Schonwiese,¨ C.-D.: Praktische Statistik fur¨ Meteorologen und Ge- gen in Nederland, Matrijs, Utrecht, 1995. owissenschaftler, Borntraeger, Berlin, 1992. Weikinn, C.: Quellentexte zur Witterungsgeschichte Europas von Sturm, K., Glaser, R., Jacobeit, J., Deutsch, M., Brazdil,´ R., and der Zeitwende bis zum Jahre 1850, Akedemie-Verl., Berlin, 1–4, Pfister, C.: Floods in Central Europe since AD 1500 and their 1958–1963. Relation to the Atmospheric Circulation, PGM, 148 (6), 18–27, Woude, J. D. van der: Holocene palaeoenvironmental evolution of a 2001. perimarine fluviatile area, Vrije Universiteit Amsterdam, Thesis, Tetzlaff, G., Borngen,¨ M., and Raabe, A.: Das Jahrtausend- 1981. hochwasser von 1342 und seine meteorologischen Ursachen,